English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We consider the problem of recovering a single person's 3D human mesh fromin-the-wild crowded scenes. While much progress has been in 3D human meshestimation, existing methods struggle when test input has crowded scenes. Thefirst reason for the failure is a domain gap between training and testing data.A motion capture dataset, which provides accurate 3D labels for training, lackscrowd data and impedes a network from learning crowded scene-robust imagefeatures of a target person. The second reason is a feature processing thatspatially averages the feature map of a localized bounding box containingmultiple people. Averaging the whole feature map makes a target person'sfeature indistinguishable from others. We present 3DCrowdNet that firstlyexplicitly targets in-the-wild crowded scenes and estimates a robust 3D humanmesh by addressing the above issues. First, we leverage 2D human poseestimation that does not require a motion capture dataset with 3D labels fortraining and does not suffer from the domain gap. Second, we propose ajoint-based regressor that distinguishes a target person's feature from others.Our joint-based regressor preserves the spatial activation of a target bysampling features from the target's joint locations and regresses human modelparameters. As a result, 3DCrowdNet learns target-focused features andeffectively excludes the irrelevant features of nearby persons. We conductexperiments on various benchmarks and prove the robustness of 3DCrowdNet to thein-the-wild crowded scenes both quantitatively and qualitatively. The code isavailable at https://github.com/hongsukchoi/3DCrowdNet_RELEASE.

Learning to Estimate Robust 3D Human Mesh from In-the-Wild Crowded Scenes